Hi All,
I am planning to run another RNAseq experiment to add additional samples to compare to my first experiment.
My first experiment had 3 controls and 4 affected patient samples. However interestingly the 4 patient samples clustered into two distinct groups of 2. To try and add to these groups and understand their significance we are going to add additional samples, including RNA from the cells of original affected patients but of later passage number, to see how they group and compare with those affected samples run in the first experiment.
We are planning to use new control (unaffected) RNA in this second round of experiment from distinct patients to those used in the first experiment... However
My question is, should we aleternatively plan to include control RNA that are from the same control patients as in the first experiment? Will this help us in removing batch affects and correlating these samples more closely in the later analysis of these two experiments, since we know then that the controls are the same samples and therefore should have the same expression pattern?
Many thanks in advance for all your help! -N
Are the 3 controls in your first experiment, 3 different persons? If yes, you can add new controls to your next experiment and use batch correction in your design with e.g., edgeR or limma voom.
Yes 3 different individual biological replicates.
You can treat all control replicates from 2 batches as one group, taking batch as a factor in your design.
Could I use the same controls in the second experiment as I have in the first? or would this not be advised? (my RNA extractions for the new control samples havent gone to plan)
I wouldn't do that, if you add more biological replicates
nof controls gets greater, which is always good for statistics. If you add batch info in your model the batches get corrected.